def check_adjoint(K, Kadj, K_input_shape, Kadj_input_shape):
'''
Checks if the operators K and Kadj are actually adjoint of eachother, i.e if
< K(x), y > = < x, Kadj(y) >
'''
x = np.random.rand(*K_input_shape)
y = np.random.rand(*Kadj_input_shape)
err = abs(dot(K(x), y) - dot(x, Kadj(y)))
return err
def conjugate_gradient_tv(K, Kadj, data, Lambda, n_it, mu=1e-4, return_energy=True):
'''
Conjugate Gradient algorithm to minimize the objective function
0.5*||K*x - d||_2^2 + Lambda*TV_mu (x)
where TV_mu is the Moreau-Yosida regularization of the Total Variation.
K : forward operator
Kadj : backward operator, adjoint of K
data: acquired data
Lambda : parameter weighting the TV regularization
mu : parameter of Moreau-Yosida approximation of TV (small positive value)
n_it : number of iterations
'''
x = 0*Kadj(data) # start from 0
grad_f = -Kadj(data)
grad_F = grad_f + Lambda*grad_tv_smoothed(x, mu)
d = -np.copy(grad_F)
if return_energy: en = np.zeros(n_it)
for k in range(0, n_it):
grad_f_old = grad_f
grad_F_old = grad_F
ATAd = Kadj(K(d))
# Calculate step size
alpha = dot(d, -grad_F_old)/dot(d, ATAd)
# Update variables
x = x + alpha*d
grad_f = grad_f_old + alpha*ATAd # TODO: re-compute gradient every K iterations to avoid error accumulation
grad_F = grad_f + Lambda*grad_tv_smoothed(x,mu)
beta = dot(grad_F, grad_F - grad_F_old)/norm2sq(grad_F_old) # Polak-Ribiere
if beta < 0:
beta = 0
d = -grad_F + beta*d
# Energy
if return_energy:
fid = norm2sq(K(x)-data)
tv = tv_smoothed(x,mu)
eng = fid+Lambda*tv
en[k] = eng
if (k % 10 == 0): # TODO: more flexible
print("%d : Energy = %e \t Fid = %e\t TV = %e" %(k, eng, fid, tv))
# Stoping criterion
if np.abs(alpha) < 1e-15: # TODO : try other bounds
print("Warning : minimum step reached, interrupting at iteration %d" %k)
break;
if return_energy: return en, x
else: return x
def conjugate_gradient(K, Kadj, data, n_it, return_energy=True):
'''
Conjugate Gradient algorithm for least squares fitting :
0.5*||K*x - d||_2^2
K : forward operator
Kadj : backward operator, adjoint of K
data: acquired data
n_it : number of iterations
'''
x = 0*Kadj(data) # start from 0
grad_f = -Kadj(data)
d = -np.copy(grad_f)
if return_energy: en = np.zeros(n_it)
for k in range(0, n_it):
grad_f_old = grad_f
ATAd = Kadj(K(d))
# Calculate step size
alpha = dot(d, -grad_f_old)/dot(d, ATAd)
# Update variables
x = x + alpha*d
grad_f = grad_f_old + alpha*ATAd # TODO: re-compute gradient every K iterations to avoid error accumulation
beta = dot(grad_f, grad_f - grad_f_old)/norm2sq(grad_f_old) # Polak-Ribiere
if beta < 0:
beta = 0
d = -grad_f + beta*d
# Energy
if return_energy:
eng = norm2sq(K(x)-data)
en[k] = eng
if (k % 10 == 0): # TODO: more flexible
print("%d : Energy = %e" %(k, eng))
# Stoping criterion
if np.abs(alpha) < 1e-15: # TODO : try other bounds
print("Warning : minimum step reached, interrupting at iteration %d" %k)
break;
if return_energy: return en, x
else: return x
